Brown Center Report 2003 - Brookings2016/06/10 · This is the fourth edition of the Brown Center Report on American Education. The report premiered in the fall of 2000, as the presidential

The 2003 Brown Center Report on American Education:

HOW WELL ARE AMERICAN STUDENTS LEARNING?

With special sections on homework, charter schools, and rural school achievement

ABOUT BROOKINGSThe Brookings Institution is a private nonprofit organizationdevoted to research, education, and publication on importantissues of domestic and foreign policy. Its principal purpose is to bring knowledge to bear on current and emerging policyproblems. The Institution maintains a position of neutrality on issues of public policy. Interpretations or conclusions inBrookings publications should be understood to be solely those of the authors.

BROWN CENTER ON EDUCATION POLICY

Established in 1992, the Brown Center on Education Policy

conducts research on topics in American education, with

a special focus on efforts to improve academic achievement

in elementary and secondary schools. For more information,

see our website, www.brookings.edu/browncenter.

To order copies of this report, please call 800-275-1447, fax 202-797-2960, e-mail [email protected], or visit online at www.brookings.edu.

This report was made possible by the generous financial support of The Brown Foundation, Inc., Houston.

The 2003 Brown Center Report on American Education:

HOW WELL ARE AMERICAN STUDENTS LEARNING?With special sections on homework, charter schools, and rural school achievementOctober 2003

Volume I, Number 4

by:

TOM LOVELESS

Director, Brown Center on

Education Policy

Copyright ©2003 byTHE BROOKINGS INSTITUTION1775 Massachusetts Avenue, NW Washington, D.C. 20036www.brookings.edu

All rights reserved

TABLE OF CONTENTS

3 Introduction

PART I

5 The Nation’s Achievement

PART II

16 Do Students Have Too Much Homework?

PART III

27 Charter Schools: Achievement, Accountability, and the Role of Expertise

36 Endnotes

Research assistance by:

PAUL DIPERNA

Brown Center on Education Policy

JOHN COUGHLAN

Brown Center on Education Policy

THE 2003 BROWN CENTER REPORT ON AMERICAN EDUCATION

This is the fourth edition of the Brown Center Report on American Education.

The report premiered in the fall of 2000, as the presidential campaign

between George W. Bush and Al Gore drew to an exciting finish. This year’s

report is published as Democratic candidates vie for their party’s nomination

for president. Education will figure prominently in the 2004 election, as it did

in the last. President Bush will herald No Child Left Behind (NCLB) as the

most important federal education legislation in a generation. Democrats will

argue that without ample funding the law is an empty promise. They will

point out that the Bush administration put the brakes on federal education

spending just as NCLB was getting off the ground.

The partisan debate over how to improve American education has been

following the same script for the past four years. Republicans stress holding

schools accountable and make concessions on funding increases. They seek

to neutralize the Democrats’ traditional strength with voters on education

issues. Democrats stress the need for providing more resources to schools

and make concessions on accountability provisions. They seek to present

education as a key plank in their party’s domestic agenda for the country.

This year’s Brown Center Report examines several issues that are important

to No Child Left Behind and ongoing efforts to improve American schools.

The first section of the report analyzes the latest data on student

achievement and asks how the nation’s students are doing in reading and

mathematics. Achievement in rural schools receives a closer look.

The second section is a study of homework. Conventional wisdom is

that higher academic standards, a bedrock of No Child Left Behind, have

The Brown Center Report on American Education 3

4 The Brown Center Report on American Education

driven up the amount of students’ homework. Stories of tired, over-worked

kids abound. After examining several different sources of data on the topic,

the study concludes that virtually no evidence exists that homework has

increased in recent years, nor that the homework load has become—or ever

was—overwhelming. The stories of children laboring under onerous

amounts of homework appear to feature a small proportion of children who,

though their predicaments are real, are not typical.

The third section of the report presents a follow-up of last year’s study

on charter schools. This year’s study examines charters’ test scores, with a

special focus on achievement in conversion charters, schools that were pre-

viously regular public schools and converted to charter status, and charters

managed by educational management organizations (EMOs), professional

management firms. Both types of charter school can lay claim to a particular

form of expertise. When a regular public school converts to a charter school,

the most talented and experienced teachers and administrators usually stay

on board. The very existence of educational management organizations is

based on the premise that expert managers, who are usually not educators

and come from the private sector, can employ their leadership skills to make

schools more productive.

Readers should be informed that the author of the Brown Center Report,

Tom Loveless, has had several professional affiliations with charter schools.

These associations cannot change the data on which the following analysis is

based, but they might have, in ways unknown, influenced the choice of

questions in the research or the interpretation of the findings. The data are

downloadable from the Brown Center website (www.brookings.edu/brown-

center) for those who would like to conduct their own analysis.

The Brown Center Report on American Education

Part

ITHE NATION’SACHIEVEMENT


Part

I


Part I The Nation’s Achievement

THIS SECTION OF THE REPORT EXAMINES FEDERAL AND STATE

assessment data to determine how well American students are

learning reading and mathematics. In the spring of 2003, the fed-

eral government released scores from the 2002 National Assessment of

Educational Progress (NAEP). As the Brown Center Report went to press,

NAEP scores from 2003 were being readied

for release, making it the first time that

comparable NAEP tests had been given in

consecutive years. Annual NAEP testing

reflects the intention of the No Child Left

Behind Act (NCLB) that the NAEP serve as

an independent monitor of state tests.

Accountability systems are based on how

schools perform on state tests. The NAEP

should, ideally, indicate whether states make

their tests too easy or otherwise report

inflated scores.

The NAEP program has two tests—

the main and the trend. The 2002 results

were derived from the main NAEP, the fifth

time the test has been given in reading since

1990. Main NAEP scores in math were last

reported in 2000. The trend NAEP, a test first

administered in 1969 in science, 1970 in

reading, and 1972 in math, will be given

again in 2004. Despite the confusion caused

by using two tests, NAEP is valuable for

providing periodic snapshots of the nation’s

academic achievement based on the perfor-

mance of large, randomly selected, nationally

representative samples of students on tests of

academic skills.

NAEP Gains by Subject and GradeTables 1-1 and 1-2 present the latest NAEP

data. Scores are broken out by subject area

and grade level. Significant findings are evi-

dent along these two dimensions. Math scores

continue to improve at a quicker pace than

reading scores. Math scores at the fourth

grade level rose by approximately .47 stan-

dard deviations from 1992 to 2000, an

extraordinary increase on any achievement

test—about one and one-fourth grade levels

of knowledge. Almost one year’s gain was

made from 1990 to 1996 alone! The gains

made by eighth and twelfth graders have been

less pronounced but are still significant.1

The reading results are quite different.

The 2002 scores are flat (see Table 1-2).

Reading proficiency has changed very little

since 1992. The languid progress in reading is


surprising given the emphasis placed on liter-

acy instruction by policymakers in the 1990s.

The federal government passed the Reading

Excellence Act of 1998, and a long, bruising

battle over phonics legislation transpired

in dozens of states.2 Reading performance

has barely budged.

Younger students are making larger

strides in academic performance than older

students. Twelfth grade reading scores have

actually declined. The loss of 5 scale score

points in that grade level since 1992 is

not only statistically significant, but also sig-

nificant in the real world. One-fourth of

twelfth graders scored below a basic level as

defined by NAEP performance standards.

That means about 700,000 students in the

final year of high school do not have the basic

reading skills that are required for meaningful

employment or success in higher education.

What explains the poor reading perfor-

mance by twelfth graders? No one knows for

sure. High school seniors may not take the

NAEP test seriously. An analysis of responses

on the 1996 NAEP uncovered evidence that

more than a quarter of twelfth graders

engaged in off-task behaviors—leaving ques-

tions blank, doodling in the margins, giving

the same answer on a long string of items, and

the like.3 How students perform on NAEP

does not count for anything, and the test’s low

stakes may hinder its ability to assess what

students really know. Determining whether

off-task test behaviors have increased in recent

years would help explain their role in affecting

trends in test scores. The fact that twelfth

grade NAEP scores in math have not similarly

declined casts doubt on the explanation.

Other theories for the disappointing

reading scores lay blame on high schools

themselves. A federal commission singled out

the senior year in American high schools as

an academic wasteland, the typical student’s

day filled with non-academic classes, study

halls, part-time jobs, and extra-curricular

activities.4 Early admission to college robs

even the most studious seniors of the incen-

tive to work hard. Many top students are noti-

fied that they have been accepted to college

as early as December, making the second half

of the school year irrelevant to their future

plans. High school seniors take so many tests

that count—finals, Advanced Placement tests,

graduation exams—that tests such as the

NAEP risk being seen as unimportant.

State Test ResultsThe Brown Center Report has been monitor-

ing state test scores for the past four years.

State test results now take on heightened

importance. Under No Child Left Behind,

state tests are the instrument by which

schools will be held accountable for produc-

ing student learning. Tables 1-3 and 1-4

summarize the latest results, compiled from

tests given in 2002. Data were obtained from

forty-nine states. The analysis is confined to

The languid progress

in reading is surprising

given the emphasis

placed on literacy

instruction by policy-

makers in the 1990s.

Reading Main NAEP Scores, 1992–2002(average scale scores)

Grade 4

Grade 8

Grade 12

1992

217

260

292

1994

214

260

287

1998

217

264

291

2000

217

—

—

2002

219

264

287

1990-2002Change

+2

+4

–5

Change in SD Units

+0.06

+0.11

–0.15

Table

1-2

Source: Standard Deviations in 1992 were: Grade 4, 36 points; Grade 8, 36 points; Grade 12, 33 points.

As measured by the National Assessment of Educational Progress (NAEP), for grades 4, 8, and 12. NAEP data are expressed as scale scores, ranging from 0 to 500.

Math Main NAEP Scores, 1990–2000(average scale scores)

Grade 4

Grade 8

Grade 12

1990

213

263

294

1992

220

268

299

1996

224

272

304

2000

228

275

301

Table

1-1

Source: Standard Deviations in 1990 were: Grade 4, 32 points; Grade 8, 36 points; Grade 12, 36 points.

As measured by the National Assessment of Educational Progress (NAEP), for grades 4, 8, and 12. NAEP data are expressed as scale scores, ranging from 0 to 500.

1990-2000Change

+15

+12

+7

Change in SD Units

+0.47

+0.33

+0.19



states using the same test in the same subject

area and grade level in 2001 and 2002. Since

states assess achievement in various grade

levels, the number of states reporting at any

particular grade varies.

Table 1-3 shows the percentage of

states reporting an improvement, no change,

or a decline in mathematics from 2001 to

2002. Table 1-4 provides the same informa-

tion for reading. The same patterns noted

in the NAEP data are apparent here, with

more states gaining in math than reading

and more states gaining at younger grades

than older grades. In math, more states

reported gains than losses in grades 4, 5,

and 8. Only at tenth grade did declining

states outpace advancers (by a margin of

eight to six, with five states reporting no

change). In reading, more states reported

gains in grades 4 and 5 than reported

declines (see Table 1-4). The fifth grade

advance is small—eight states advancing to

six declining. In grades 8 and 10, declining

states outpaced gainers.

How do the latest state scores compare

with previous years? Is student achievement

in the U.S. improving or deteriorating? The

direction and magnitude of national

achievement must be discerned from several

disparate snapshots. To evaluate national

performance in previous years, the Brown

Center Report tracked over time the per-

centage of states with rising test scores, as

the first rows of Tables 1-3 and 1-4 report

for 2002. The problem with that approach

is that it does not adequately account for

states with falling scores. This year, Brown

Center researchers devised a new statistic

to serve as an indicator of momentum in state

achievement tests.

State Test Results(2001–2002)

Math

Improvement

No Change

Decline

Total

�

��

�

Grade 4

16 (50%)

8 (25%)

8 (25%)

32 states

Improvement

No Change

Decline

Total

�

��

�

Grade 5

11 (55%)

3 (15%)

6 (30%)

20 states

Improvement

No Change

Decline

Total

�

��

�

Grade 10

6 (32%)

5 (26%)

8 (42%)

19 states

Improvement

No Change

Decline

Total

�

��

�

Grade 8

15 (38%)

11 (28%)

13 (33%)

39 states

Table

1-3

0.01

0.59

0.32

0.70 0.68

0.45

0.20

0.32

-1.0

0.0

0.2

0.2

0.4

0.4

0.6

0.6

0.8

0.8

1.0

1999 2000 2001 2002

Reading

Math

Fig

1-1State tests show math gains outpacing gains in reading…

Achievement Momentum Index by subject, 1999–2002

But improvement is slowing in both subjects.

Test ScoreGains

Test ScoreLosses

Source: Test data obtained from 49 states (and the District of Columbia) that administered the same achievement test in consecutive years.

Source: Data obtained from 49 states(and the District of Columbia) thatadministered the same achievementtest in consecutive years.


Achievement Momentum IndexThe index is simple to understand. For any

particular year, the index ranges from –1.00

to +1.00. Each state’s contribution to the

index is weighted to represent its share of the

nation’s student population. The calculations

include reading and math scores for grades 4,

8, and 10. For states reporting higher test

scores than the previous year, their popula-

tion weights are multiplied by +1.00. For

states reporting lower scores, their popula-

tion weights are multiplied by –1.00. States

reporting the same scores are set at 0.00. The

products are summed to create the index.

Interpreting the index is straightfor-

ward. A positive momentum score means test

scores are going up. A negative score indi-

cates they are going down. And a 0.00 means

that scores are unchanged or that the num-

bers of students in improving and declining

states are about equal. If every state reports

test scores that have improved from the previ-

ous year, the momentum index will read

+1.00. If every state reports test scores that

have declined, the index will read –1.00. The

closest corollary to the index is a corpora-

tion’s annual statement of profits or losses,

which compares current performance with

performance in the previous year. Since the

index is population weighted, a gain in a large

state like California can offset losses in several

small states and vice versa.

Figures 1-1 and 1-2 display the

momentum indices from 1999 to 2002.

Figure 1-1 reports by subject area, with

scores averaged across grades 4, 8, and 10.

And Figure 1-2 reports by grade level, with

math and reading scores averaged at each

grade. As shown in Figure 1-1, scores in

reading and math remained positive in 2002

but less so than in recent years. Progress

has slowed. The highest reading scores were

recorded in 1999, the first year the Brown

Center began collecting state test data. Math

scores peaked in 2000. Nevertheless, both

momentum indices are signaling that

students continue to improve, albeit more

slowly, in reading and math.

State Test Results(2001–2002)

Reading

Improvement

No Change

Decline

Total

�

��

�

Grade 4

16 (48%)

6 (18%)

11 (33%)

33 states

Improvement

No Change

Decline

Total

�

��

�

Grade 5

8 (40%)

6 (30%)

6 (30%)

20 states

Improvement

No Change

Decline

Total

�

��

�

Grade 10

5 (36%)

3 (21%)

6 (43%)

14 states

Improvement

No Change

Decline

Total

�

��

�

Grade 8

12 (32%)

11 (30%)

14 (38%)

37 states

Table

1-4

Source: Data obtained from 49 states(and the District of Columbia) thatadministered the same achievementtest in consecutive years.

The National Assessment of Educational

Progress (NAEP) is commonly referred to

as the Nation’s Report Card. Since 1969,

it has been the only nationally representa-

tive and continuing assessment of what

America’s students know and can do in

academic subject areas. The number

of students selected for a NAEP national

sample for any particular grade and sub-

ject is 7,000 or more.

There are three NAEP test types:

(1) the main NAEP gauges national

achievement while also reflecting current

practices in curriculum and assessment,

(2) the long-term trend NAEP allows

reliable measurement of change in national

achievement over time, and (3) the state

NAEP measures achievement of students

in participating states. These assessments

use distinct data collection procedures

and separate samples of students.

Since 1990, the main and state math

tests have been governed by a framework

reflecting recommendations of the National

Council of Teachers of Mathematics

(NCTM). The long-term trend test consists

of essentially the same items and test

procedures used in 1973.

Introduction to NAEP


Figure 1-2 displays academic progress

by grade. Scores at the fourth grade level

remain strong, with 2002 the best year of

the four. Eighth and tenth graders continue

to tally test score gains, but momentum

has tailed off since 2000. Here, the positive

signal for eighth and tenth grades in 2002

is contrary to that given above by Tables 1-3

and 1-4 for the same year. The momentum

index was driven into positive territory

by gains in California and Texas, two

large states.

The latest federal and state test

results give a clear indication of how well

American students are learning. Student

achievement continues to improve in both

reading and math, but at a slower pace than

three or four years ago. Math gains outpace

reading gains, and elementary grade children

are improving more than middle and high

school students. The analysis now shifts to

an often overlooked population in discus-

sions of test scores.

Rural SchoolsRural schools are America’s forgotten educa-

tional institutions. Educational reformers

generate moral energy by championing the

cause of urban education. This focus also

pays political dividends. Federal and state

officials invariably craft education policies

with urban schooling as a central concern.

It is no accident that education’s two most

enduring policy debates—concerning

parental choice and school finance—have

been largely contested over the effects

of policy on children in big cities.

The online archives of Education Week,

the national journal of record in the field of

education, offer a yardstick for measuring

Part I The Nation’s Achievement Rural schools are

America’s forgotten

educational institu-

tions. Educational

reformers generate

moral energy by

championing the cause

of urban education.

Source: Test data obtained from 49 states (and the District of Columbia) that administered the same achievement test in consecutive years.

-1.0

0.0

0.2

0.2

0.4

0.4

0.6

0.6

0.8

0.8

1.0

Test Score Gains

Test Score Losses

0.56

0.550.55

0.01

0.63

0.72 0.69

0.66

0.61

0.26

0.20

0.75

1999 2000 2001 2002

8th Grade

4th Grade

State tests show 4th graders making greater gains than older students…

Achievement Momentum Index by grade, 1999–2002

But the progress of 10th graders has slowed for two consecutive years.

10th Grade

Fig

1-2


media coverage of education in different

types of communities. Consider the number

of times “urban, ” “rural, ” and “suburban”

were mentioned in Education Week from 2000

through 2002. “Rural” appeared 425 times,

“suburban” 346 times, and “urban” 916

times.5 Of all such mentions, the word

“rural” appeared 25% of the time, “suburban”

21%, and “urban” 54%.

Note that suburban schools received

even less attention than rural schools.

That is probably because they are assumed

to be amply endowed with resources and

to be at least adequate, if not excellent, in

quality. Polls show that parents and the gen-

eral public are overwhelmingly satisfied with

suburban schools. But what about rural

schools? They are rarely mentioned in the

public debate over school reform. Are they as

successful as suburban schools? What kinds

of students attend rural schools? Do they

learn as much as students in other settings?

Characteristics of Rural SchoolsRural schools are not monolithic. They look

very different in different regions of the

country, serve different kinds of students,

operate in different economic and social con-

texts, and encounter different challenges.

Table 1-5 compares urban, suburban, and

rural schools on several characteristics.

However, as one reads the ensuing discus-

sion, it should be remembered that for every

characteristic that is exhibited by most rural

schools, pockets of rural schools exist with

contrasting profiles.

Rural schools enroll only 27% of the

nation’s students but occupy 42% of school

buildings. As a consequence, the average

rural school is quite small. It serves only 392

students, which is about 40% smaller than

the average urban or suburban school.

Compared to the typical urban school, the

typical rural school has a smaller proportion

of students who qualify for free and reduced

lunch, a measure of poverty. And rural

schools serve a much larger proportion of

white students than urban or suburban

schools. The black and Hispanic populations

are sparse in rural schools. Only 8% of rural

students are black and 7% are Hispanic,

compared to 33% who are black and 24%

who are Hispanic in urban schools.

Rural schools receive less funding than

most schools, about 13% less than urban

schools ($5,734 vs. $6,575). In part, the

funding shortfall may be attributable to a

greater reliance by rural schools on state

funds. Rural schools receive 53% of their

revenues from states, compared to 44% for

suburban schools. Urban and suburban

districts tap local property wealth to boost

revenues. A lower cost of living also

undoubtedly plays a part. Recent calcula-

tions by the Economic Research Bureau

estimate that the cost of living in rural areas

is 16% less than in urban areas.6

The problems rural schools face in

gaining sufficient financial resources and

well-trained teachers have been documented

Table

1-5Characteristics of Urban, Suburban, and Rural Schools

% of Nation’s Students

% of Nation’s Schools

Mean School Size

% Free/Reduced Lunch

% White

% Black

% Hispanic

% Asian

% Native American

Per Pupil Expenditure

% of Revenue from State

Urban

30

24

663

57

37

33

24

5

1

$6,575

49

Suburban

43

34

665

32

66

13

15

5

1

$6,229

44

Rural

27

42

392

39

80

8

7

1

4

$5,734

53

Source: National Center for Educational Statistics (NCES), “Navigating Resources for Rural Schools: PublicElementary and Secondary Students, Schools, Pupil/Teacher Ratios, and Finances, by Type of Locale: 1998and 1999”, (NCES 2001); National Center for Educational Statistics (NCES), 2002 Common Core of Data.


by The Rural School and Community Trust.7

Rural schools may also lack other types of

education resources. Some of the shortages

stem from rural schools being exceptionally

small. Small schools offer fewer curriculum

options and provide fewer opportunities for

high achievers to enroll in advanced place-

ment or International Baccalaureate pro-

grams. Extracurricular activities are limited.

Even rural schools that possess the facilities

to offer a specialized program—say, an

advanced placement chemistry class—may

have a difficult time staffing it with a well-

qualified teacher.

Choice and Public SupportParents exercise less choice in rural areas.

Only 15.9% of students in rural areas attend

public or private schools of choice, com-

pared to 28.7% in non-rural areas. It is diffi-

cult to say whether this pattern is the result

of short supply or lack of demand. Rural

parents have fewer alternatives to conven-

tional public schools, either in the private

sector or in the public sector through mag-

net or charter schools. Little more than 5%

of rural students attend private schools,

about half of the national percentage.8

Rural parents are satisfied with public

schools, and the support for public schools

extends into the community. In the 2002 Phi

Delta Kappa/Gallup Poll, rural residents gave

their schools higher grades than either sub-

urban or urban residents. In awarding letter

grades to local public schools, 53% of rural

residents assigned an A or B, compared to

51% of suburban residents and 32% of

urban residents. In a survey of students con-

ducted by Metropolitan Life, 59% of rural

students described parental and community


212

223 220

Grade 4200

250

275

225

300

268 266

258

Grade 8

290285

Grade 12

284

Suburban

Rural

Urban

Rural 4th graders achieve at levels close to that of suburban students.

By 12th grade, rural students’ performance resembles that of urban students.

2002 NAEP reading average scale score by location

Source: As measured by the National Assessment of Educational Progress (NAEP), for grades 4, 8, and 12. NAEP data are expressed as scale scores, ranging from 0 to 500.

Fig

1-3

A z-score is the distance fromthe mean expressed in standarddeviation units. A z-score of 0.00is, by definition, the mean. A z-score of 1.00 indicates that aschool’s test scores are onestandard deviation above themean, approximately at the 84thpercentile. A z-score of -1.00 puts a school at approximatelythe 16th percentile.

What are z-scores?


support for their schools as excellent or pret-

ty good. Approximately 51% of suburban

students and 44% of urban students offered

such a positive assessment of support for

their schools. Compared to suburban and

urban students, rural students were also

more likely to give their teachers high grades

on such critical aspects of teaching as under-

standing subject matter, treating students

with respect, helping students, and main-

taining control and discipline in the class-

room.9 A recent study by Richard Ingersoll

found that rural schools have lower teacher

turnover than other schools nationally.

Relatively stable teaching staffs may con-

tribute to the confidence that local residents

have in rural schools.10

AchievementHow do rural students perform on state

achievement tests? Table 1-6 displays 2002

test data from fourteen states. Reading and

math scores were combined for grades 4, 8,

and 10. Because the states give different

tests and report scores on different scales,

Brown Center researchers converted school

test scores to z-scores, which designate

where schools fall on the distribution of all

scores within a state. By definition, z-scores

have a mean of 0.00 and standard deviation

of 1.00. For each of the fourteen states,

the average z-score of rural schools is report-

ed in the final column of Table 1-6.

Rural schools are generally perform-

ing quite well, scoring above average in all

states except for Arizona, Minnesota, and

South Carolina. Solid test scores may be a

key reason why rural schools enjoy broad

community support. It is impossible for

these data to tell us much about why rural

students do so well. Almost all of the test

score advantage washed out when regres-

sion analyses controlled for schools’ racial

composition and percentage of students in

poverty. Those demographic statistics are

also listed in the table.

Federal NAEP data provide another

source of data on rural school achievement.

Figure 1-3 displays the 2002 NAEP reading

scores by grade level and location. At fourth

grade, students in rural schools achieve at

levels similar to students in suburban

schools, scoring one point above the nation-

al average of 219. By twelfth grade, rural stu-

dents score about the same as urban

students, two points below the national aver-

age of 287. This slight slippage in rural stu-

dents’ scores from fourth to twelfth grades

shows up in previous NAEP tests—in read-

ing as well as other subjects. Rural students

appear to do better on achievements tests in

elementary school than in high school.

A more profound cause for concern

can be found in data gathered by the

National Center for Education Statistics in

1993-94. First, the good news (see Figure

1-4). In a survey of schools with a twelfth

Achievement and Socioeconomic Characteristics for Rural Schools(2002)

State

Arizona

California

Colorado

Florida

Georgia

Louisiana

Massachusetts

Michigan

Minnesota

North Carolina

Pennsylvania

South Carolina

Texas

Wisconsin

Number of RuralSchools

464

1,106

615

682

990

620

381

1,412

833

1,215

1,267

573

2,344

1,103

% in State

27.3

14.4

40.1

22.8

50.8

45.0

21.1

40.4

51.7

56.6

40.3

54.5

34.9

54.1

% FreeLunch

in RuralSchools

34.1

44.3

30.8

44.1

47.9

62.5

13.1

33.4

32.4

46.1

25.9

59.6

47.0

25.1

StateAverage

43.3

48.1

31.2

47.4

50.0

64.0

27.0

36.6

31.4

45.2

29.9

54.9

48.1

28.2

% Non-white

in Rural Schools

53.6

37.1

21.5

31.8

34.9

40.8

5.0

7.8

8.8

35.1

4.6

49.9

39.5

6.0

Table

1-6

NOTE: Arizona’s poverty data are from 2001. Test scores collected from each state’s department of education.

Source: Demographic data from the National Center for Educational Statistics (NCES),2002 Common Core of Data.

* ρ < .05, two-tailed test of z-score = 0

StateAverage

50.3

59.8

31.6

47.4

47.7

52.1

23.8

25.5

17.9

41.4

20.6

48.3

55.5

17.4

MeanZ-score in Rural Schools

–0.23*

0.11*

0.20*

0.16*

0.10*

0.05

0.28*

0.19*

–0.07

0.06

0.16*

–0.21*

0.16*

0.20*

Schools Poverty Race Achievement


grade, rural schools reported a low drop-out

rate in the senior year. Of students who

started their senior year in the fall, 94.6% in

rural schools graduated in June, compared

to 92.5% in suburban schools and 89.7% in

urban schools. However, the bad news is

that the promising graduation rate does not

convert to a high college application rate.

Only 54.3% of rural seniors apply to col-

lege, compared to 56.5% of urban and

61.6% of suburban seniors. Put another

way, for every 100 suburban high school

graduates, 67 students apply to college. The

comparable figure for urban students is 63

students. For rural schools, it is 57 students.

Nationwide, tens of thousands of

rural students are slipping through the cracks

in the transition from high school to

college. As reported by a recent University

of Chicago study, college matriculation

rates may be pushed down by students want-

ing to stay close to home after high school. In

many rural areas, post-secondary institutions

simply may not exist. Moreover, rural stu-

dents frequently believe that local communi-

ty colleges offer the same college experience

as far-away four-year institutions.11

Summary and DiscussionThis section of the Brown Center Report ana-

lyzed the most recent data on reading and

math achievement. Overall, test scores are

rising. Students continue to make solid

progress in mathematics, but reading

achievement is, at best, only inching for-

ward. Younger students are making greater

academic gains than older students.

Students in rural schools score well

above average on state tests of academic

achievement. Rural students who begin their


94.6

54.3

Rural0%

20%

40%

60%

80%

100%

61.6

92.5

Suburban

56.5

Urban

89.7

% Graduating from High School

% Applying to College

Rural 12th graders graduate from high schools in large numbers…

But they are under- represented among college applicants.

Percentage of students

Source: Data obtained from the NCES website “Navigating Resources for Rural Schools”: http://nces.ed.gov/surveys/ruraled

Fig

1-4

Nationwide, tens of

thousands of rural

students are slipping

through the cracks

in the transition from

high school to college.


senior year in high school are also more likely

to graduate in June than seniors in urban or

suburban schools. But the high test scores

and graduation rates are not reflected in high

rates of applying to college. Rural students

are less likely to apply to college than urban

and suburban students. The result is the loss

of potential for rural youth and a loss of tal-

ent for the nation’s colleges and universities.

Recent studies have highlighted several

unique challenges facing rural schools.

Schools in rural areas find recruiting new

teachers difficult. Rural states have many

teachers who teach several subjects, making

it hard to meet the requirement of No Child

Left Behind that teachers major in the aca-

demic subject they teach or pass a test

demonstrating mastery of the subject. The

stipulation that parents whose children

attend failing schools be given the choice

of attending another school nearby is a

daunting endeavor for some rural schools.

On a visit to Alaska in May 2003, Secretary

of Education Rod Paige visited a public

school located on an island where the closest

public school is 164 miles away—across

the Bering Sea.12

These are legitimate problems, and

policymakers should make every effort to see

that they are solved. Policymakers should

also consider programs designed to boost the

rate of rural students applying to colleges

and universities. The academic futures of

rural youth should not end upon completion

of high school.


Part

IIDO STUDENTSHAVE TOO MUCHHOMEWORK?



Part II Do Students Have Too Much Homework?

Almost everything in this story is wrong.

A wealth of data exists on the topic of

homework. The evidence paints a more

placid picture than the stories in the press.

The most reliable data support the follow-

ing conclusions: 1) the typical student, even

in high school, does not spend more than

an hour per day on homework, 2) the

homework load has not changed much

since the 1980s, 3) the students whose

homework has increased in the past decade

are those who previously had no homework

and now have a small amount, 4) most par-

ents feel the homework load is about right

and, of those who would like to change it,

more parents would rather see homework

increased than decreased.

Why is it important to get the home-

work story right? Mainly because it is posi-

tively associated with student learning.

Research shows that the relationship of

homework with student achievement is

positive for both middle and high school

students and neutral for elementary school

students. The research does not prove

causality, an ever-present difficulty with

research on many educational practices.

High-achieving students in high school,

for example, may do more homework

because they enjoy studying. They take

tough classes that require a lot of work.

That does not necessarily mean that home-

work is boosting their achievement.

Low-achieving students in elementary

school, on the other hand, may do more

homework because they are struggling

to catch up. The homework is not causing

their learning problems.13

Despite these limitations, the evi-

dence is convincing enough to provide

sober guidance on the current controversy.

Excessive homework is not a common

problem. The critics of homework need to

produce some very powerful evidence

before policymakers start mandating reduc-

tions in homework or even banning it alto-

gether. To date, the evidence put forth by

homework critics has been weak.

So where has the notion of a crushing

homework burden come from? And why has

the press latched onto it with such zeal? Note

that the summary above employs terms such

as “typical student” and “most parents.”

There are children with too much home-

work. There are also parents who believe,

correctly, that their kids are overworked.

Generalizations, however, are meant to apply

generally. Anecdotes can be woven together

to create dramatic stories, but if they apply

only to a small minority of people, they

should not be construed to depict the experi-

ence of the average person. Most parents and

SEVERAL MAJOR NEWSPAPERS AND MAGAZINES HAVE RUN ARTICLES

describing a backlash against homework. The typical story is that

dramatic increases in the amount of homework are robbing American

students of their childhood, turning kids off learning, and destroying

family life. A revolution is brewing. Kids are buried in homework.

Parents are hopping mad, and they’re going to do something about it.



children are comfortable with the current

amount of homework. The amount of home-

work assigned today does not appear out of

line with the amount of homework assigned

in the past.

Yet dramatic stories abound. In 1998,

Newsweek ran an article called “Homework

Doesn’t Help,” which begins by telling the

story of Adam, whose “long, sad battle with

homework” reached a crescendo of three

hours per night in fourth grade. The article

warns that “the trend among schools to pile

on more homework, starting in kindergarten,

could backfire.” Then Time published a 1999

piece, “The Homework Ate My Family: Kids

Are Dazed, Parents Are Stressed, Why Piling

It On Is Hurting Students.” In 2003, People

produced “Overbooked: Four Hours of

Homework for a Third Grader? Exhausted

Kids (and Parents) Fight Back.” Newspapers

have contributed, too. Since 2001, feature

stories about onerous amounts of homework

have appeared in the New York Times,

Washington Post, Chicago Sun-Times, Los

Angeles Times, Seattle Times, Raleigh News and

Observer, The Tennessean, Lexington Herald

Leader, and Tampa Tribune.

Schools piling it on. Kids exhausted.

Parents fighting back. Families being eaten.

Is this the truth about homework? Let’s

examine data addressing two questions.

How much homework do today’s students

have? Has the amount changed appreciably

over time?

The Michigan StudyThe study most often cited to support the

idea that homework has gotten out of con-

trol was conducted by the Population

Studies Center at the Institute for Social

Research, University of Michigan. The

research was led by Sandra L. Hofferth and

John F. Sandberg and focused on how fami-

lies spend their time. Analyzing twenty-four

hour time diaries recorded by randomly

selected families, the researchers compared

data on the activities of children ages 3-12 in

1981 and 1997. Table 2-1 shows the average

number of minutes devoted to activities that

consume at least two hours of children’s time

per week.14

Note that the relevant response catego-

ry is “studying,” which certainly includes

homework but could also include self-initi-

ated studying or studying that parents insist

children do, irrespective of whether the

school sends work home. It ranks quite low

on the list of activities that absorb children’s

time at home. More than studying, kids are

sleeping, going to school, playing, eating,

attending to personal care, participating in

sports, going to daycare, shopping, visiting,

and working around the house. To put it in

perspective, studying ranks close to the cate-

gory “other passive leisure”—board games,

collecting baseball cards, and the like.

For all ages, the amount of weekly

time devoted to studying increased from one

hour 53 minutes in 1981 to two hours 16

minutes in 1997, an increase of 23 minutes.

That is 19 to 27 minutes per day of studying

in 1997 and an increase of about 3-5 min-

utes per day since 1981, depending on

AN ERRONEOUS VIEW FROM THE POPULAR PRESS

Schools piling it on.

Kids exhausted. Parents

fighting back. Families

being eaten. Is this the

truth about homework?

© 1999 TIME Inc., reprinted by permission.


whether weekends are counted. It does not

seem like a backbreaking load. Nor an

extraordinary increase. Time spent on per-

sonal care and sports increased much more.

Watching television decreased by four hours

per week, and some of this freed-up time

may have gone to the 23 minutes of

increased study time. That sounds like cause

for celebration, not hand wringing. Children

have not abandoned TVs. The thirteen and a

half hours that children continue to devote

to television each week is more than six

times the amount of time spent studying.

Changes in time devoted to studying

were driven by a large change for ages 6-8,

mostly first through third graders. The

average amount of study time for that age

group rose from 52 minutes a week to two

hours and 8 minutes, an increase of 76

minutes. Based on a seven day week, the

daily equivalents are about 7 minutes of

homework in 1981 and 18 minutes in 1997,

an increase of 10-11 minutes per day. For

other age groups, changes in the homework

load have been trivial.

Now we’ve reached the nub of the

matter. If one is determined to use the

Michigan data to argue that homework is

onerous, the analysis must be limited to

changes that occurred with ages 6-8. Even

for that age group, the change in study time

was less than meets the eye. The total

increase was no more than 11 minutes per

day. And the average went up primarily

because the percentage of children who had

no homework at all declined. In 1981, only a

third of children ages 6-8 spent any time at

all on studying. In 1997, about half did. So

the rising average has nothing to do with a

crushing homework burden. It is largely

attributable to children with no homework

now receiving a small amount.

A remarkable story—overlooked in

press coverage of the Michigan study—is

found in the other age groups. In every case,

fewer children were doing homework in

1997 than in 1981 (see Figure 2-1, page

20). Even among children ages 9-12, the

oldest group in the study, only 62% spent

time studying at home in 1997, down from

82% in 1981. In other words, more than

one-third of American school children ages

9-12, who are mostly attending fourth

through seventh grades, do not do any

Weekly Time Spent on Activities by Age(hours:minutes)

Sleeping

School

Television

Playing

Eating

Personal Care

Sports

Daycare

Shopping

Visiting

Household Work

Studying

Other Passive Leisure

1981 (N=61)

77:19

14:30

15:14

25:50

9:43

6:18

1:31

0:10

2:35

2:58

2:09

0:25

2:59

1997 (N=665)

76:11

12:05

13:52

17:21

9:24

8:32

4:08

7:30

3:44

3:04

2:20

0:36

2:35

***

***

***

*

1981 (N=60)

70:04

27:52

15:55

14:58

9:08

6:13

6:01

0:12

0:59

3:40

2:49

0:52

1:58

1997 (N=602)

70:49

32:46

12:54

11:10

8:05

7:53

5:13

1:33

2:38

2:48

2:07

2:08

1:33

Table

2-1

*ρ< .05, **ρ<.01, ***ρ<.001, for 1981-1997 changes

Source: Data resorted from Table 2, Sandra L. Hofferth and John F. Sandberg, “Changes in American Children’s Use of Time, 1981-1997,” Children at the Millennium: Where Have We Come From, Where Are We Going? edited by S. Hofferth and T. Owens, (Elsevier Science Ltd., 2001), pp. 193-229.

**

*

**

**

***

***

***

***

1981 (N=101)

65:36

29:02

20:01

7:24

8:13

6:21

4:51

0:18

1:57

3:48

5:18

3:22

3:24

1997 (N=851)

67:24

34:03

13:36

8:54

7:23

7:53

6:33

0:24

2:24

2:40

3:42

3:41

2:19

**

***

***

*

**

***

**

**

***

*

1981 (N=222)

70:01

24:45

17:35

14:30

8:52

6:18

4:15

0:14

1:52

3:32

3:46

1:53

2:53

1997 (N=2119)

71:07

26:48

13:29

12:12

8:13

8:05

5:25

2:57

2:53

2:50

2:49

2:16

2:11

*

**

***

***

***

***

***

**

***

*

***

Age 3-5 Age 6-8 Age 9-12 All Ages



homework. For the broader group of chil-

dren in the Michigan study, ages 3-12, half

did no homework in 1997.

The Michigan study is an example of

carefully conducted social science and the

researchers’ dispassionate reporting of

results. The authors explain that fluctua-

tions in the amount of time that children

engage in various activities should be

understood in the larger context of change

occurring in families. Compared to 1981,

children now spend less time in discre-

tionary activities such as playing and

watching television and more time in struc-

tured activities such as studying, reading,

and participating in arts and sports pro-

grams. With both parents working and

more things to accomplish during the day,

today’s families manage children’s schedules

more carefully. Nevertheless, the study’s

findings have been used to give the impres-

sion that a rising homework burden is

swamping families. “Homework Hours

Tripled Since 1980,” ran the Associated

Press headline.15

This impression does not reflect how

the Michigan study’s authors interpret their

own findings. As to whether the amount of

time studying is large or small, Hofferth and

Sandberg conclude, “While time spent

studying increased significantly between

Age 3-50%

20%

40%

60%

80%

100%

Age 6-8 Age 9-12 All Ages

1981

1997

Fewer children were doing homework in 1997 than in 1981.

Percentage of children who study—at all

The only exception is children aged 6-8, two-thirds of whom did no homework at all in 1981.

26

17

55

34

62

82

46

54

Source: Sandra L. Hofferth and John F. Sandberg, “Changes in American Children’s Use of Time, 1981-1997,” Children at the Millennium: Where Have We Come From, Where Are We Going? edited by S. Hofferth and T. Owens, (Elsevier Science Ltd., 2001), pp. 193–229, Table 1.

Fig

2-1


From Newsweek, March 30 © 1998 Newsweek, Inc. All rights reserved.Reprinted by permission.


1981 and 1997, the time spent studying is

still small, only about two hours per week,

and the increase was concentrated among

6-8 year olds.” As to the cause of the

increase among 6-8 year olds, the

researchers explain, “The main reason for

the increase in studying among 6-8 year

olds was an increase in the proportion who

did some studying at all, from one-third

to more than half.” Of the dozens of press

accounts that cited figures from the

Michigan study and were reviewed for this

Brown Center Report, not a single one

quoted how the researchers themselves

interpreted their data.16

Asking Kids How MuchHomework They HaveAs just reported, about 50% of kids have

no homework at all, which means that a

lot of zeros are going into computing the

average student’s homework load. Even

a trivial increase in the average could hide

the fact that students who once had some

homework are suddenly getting a lot more.

An excellent source of data to explore

whether this is happening is the National

Assessment of Educational Progress

(NAEP), which for two decades has been

asking a nationally representative sample

of students how much homework they

had the previous day. Table 2-2 shows

the responses for ages 9, 13, and 17 since

the early 1980s.17

The NAEP data reinforce the notion

that it is students who once had no home-

work who have experienced a change. They

now have some, but not a lot. For 9 year

olds, the question about homework was first

asked in 1984. The proportion with no

assigned homework declined by ten percent-

age points between 1984 and 1999, while

those with less than one hour, the lowest cat-

egory indicating any homework, increased

by twelve points. The same shift, although

less pronounced, is evident at age 13. Students

without assigned homework declined by six

percentage points from 1980 to 1999, and

there was a five point rise in students with

less than an hour.

Seventeen year olds’ homework

appears to have peaked in the surveys

conducted from 1984 through 1992, with

slight decreases since then. Across all three

age groups, the percentage of students

with an hour or more of homework—the

category where the horror stories should

show up—has declined since 1984. In

1999 only about one-third of students at

ages 13 and 17 had an hour or more per

Students Were Asked: How much time did you spend on homework yesterday?(percentage of students)

NoneAssigned

Did Not Do It

< 1 hr.

1-2 hrs.

> 2 hrs.

1984

36

4

41

13

6

1988

29

4

47

13

7

1992

32

4

47

12

5

1996

26

4

53

13

4

Table

2-2

1999

26

4

53

12

5

NoneAssigned

Did Not Do It

< 1 hr.

1-2 hrs.

> 2 hrs.

1980

30

6

32

24

7

1984

23

4

36

29

9

1988

17

4

37

30

11

1992

21

4

36

29

10

1996

22

5

37

27

8

1999

24

5

37

26

8

NoneAssigned

Did Not Do It

< 1 hr.

1-2 hrs.

> 2 hrs.

1980

32

12

24

23

10

1984

22

11

26

27

13

1988

21

13

28

26

12

1992

22

12

29

25

11

1996

23

13

28

24

11

1999

26

13

26

23

12

Age 9

Age 13

Age 17

NOTE: Age 9 students were not surveyed on homework until 1984.

Source: NAEP 1999 Long-term Trend Reading Summary Data Tables for ages 9, 13, and 17.



night. And combining the responses “none

assigned,” “did not do it,” and “less than

1 hr” yields the percentage of children with

less than an hour of daily homework. For

1999, the figures are 83% at age 9, 66% at

age 13, and 65% at age 17 (see Table 2-3).

Based on these statistics, it appears

that less than one hour per day is the typical

amount of homework for American children,

even for students in high school. Remember

the third grader in the Newsweek story with

three hours of homework? Or the families in

the Time article, “The Homework Ate My

Family?” Or People asking whether four

hours was too much homework for a third

grader? These tales, though very real to the

children and parents experiencing them, are

what statisticians refer to as “outliers.” They

are not representative. A whopping 83% of

9 year olds have an hour or less of daily

homework. Students with more than two

hours constitute only 5% of 9 year olds.

Reasonable people could argue that two

hours is too much homework for a third

grader, at least on a regular basis. But accord-

ing to survey data, that amount of home-

work is quite rare.

High School StudentsInterestingly, high school students are usual-

ly left out of the public discussions of home-

work. Most of the stories of schools “piling it

on” focus on young children. The NAEP data

give a hint that high school students’ home-

work load might be considered light, with

only about a third of 17 year olds having an

hour or more of daily homework. That does

not sound like a lot for students who are

within a year of graduating from high school.

How does it compare with students in other

countries? It is an extraordinarily light load.

The Third International Mathematics and

Science Survey in 1995 asked students in

their final year of secondary school how

many hours per day they spent studying or

doing homework. Of twenty nations, the

U.S. ranked near the bottom, tied for the

next-to-last position. Students in France,

Italy, Russia, and South Africa reported

spending at least twice as much time on

homework as American students.18

Even when surveys are restricted to

the top high school students in the country,

those who go on to attend college, the

homework load does not appear heavy.

Researchers at UCLA have been surveying

college freshmen nationwide annually

since 1966. The researchers began asking a

question about homework in 1987. That

year, 47% of respondents reported that they

spent more than five hours per week study-

ing or doing homework during the senior

year of high school. The figure has dropped

every year since then, hitting a record low of

34% in 2002 (see Figure 2-2). That means

two-thirds of college-bound high school

seniors have no more than one hour of

homework per night and none on week-

ends. And remember that these are students

who were preparing for college and

probably took the toughest courses that

high schools have to offer—honors,

Advanced Placement, and International

Baccalaureate.19

Two additional findings from the UCLA

survey stand out. First, at the same time stu-

dents are reporting record lows in the amount

of homework given in high school, they are

reporting record high grade point averages.

Almost half, 46%, of the college freshmen

reported graduating from high school with an

A average. Second, in a pattern remarkably

similar to the young children in the Michigan

study, college students in the UCLA survey

seem to live well-rounded lives. Socializing

with friends, working at paid employment,

and exercise and sports took up more of the

students’ time as high school seniors than

Percentage of Students ReportingLess Than an Hour of Daily Homework

Age

9 Year Olds

13 Year Olds

17 Year Olds

Percent

83%

66%

65%

Source: NAEP 1999 Long-term TrendReading Summary Data Tables for ages9, 13, and 17.

Table

2-3


studying and homework. About one-fourth of

college freshmen report that they spent more

than five hours per week partying or watching

television as high school seniors, a little less

than the one-third who spent that much time

studying (see Figure 2-3, page 24).

What Do Parents Think?Impressions matter, and despite the

large amount of empirical evidence to the

contrary, parents may think children have

too much homework. If so, that would

explain the reports of rebellions cropping

up in local areas. What do parents think

about the homework load? Public Agenda

Foundation conducted a poll on the

matter in 2000, surveying a representative

sample of parents. They were asked to eval-

uate their children’s homework load (see

Figure 2-4, page 25). Almost two-thirds,

64%, described the amount of homework

as “about the right amount,” 25% said there

was “too little homework,” and 10%

responded that there was “too much home-

work” (2% responded “don’t know”).20

These attitudes are the opposite of those

expressed by parents in the anti-homework

articles. Most parents are satisfied with the

amount of homework, and of those who

favor a change, more parents would

increase rather than decrease the amount.

Only one out of ten parents believe there is

too much homework.

Summary and DiscussionIn recent years, newspapers and magazines

have published several stories about a

growing amount of homework in the United

States. Facing mounting pressures to per-

form, schools are piling on the homework,

34%More than 5 hours

66%5 hours or less

Two-thirds of college-bound seniors do five hours or less of homework per week.

College freshmen were asked:During your last year in high school, how many hours did you spend in a typical week studying or doing homework?

Fig

2-2

Source: Linda Sax, J.A. Lindholm, A.W. Astin, W.S. Korn, and K.M. Mahoney, The American Freshman: National Norms for Fall 2002, (Higher Education Research Institute, UCLA, 2002).

Most parents are satis-

fied with the amount

of homework, and of

those who favor a

change, more parents

would increase

rather than decrease

the amount.



kids are stressed out, and parents are

demanding relief.

The stories are misleading. They do not

reflect the experiences of a majority—or even

a significant minority—of American school

children. The best social science on the topic,

which consist of studies collecting data from

nationally representative samples of students,

are in general agreement. A University of

Michigan study shows that homework for

elementary schoolchildren has increased

slightly since the early 1980s. But the

increase was due to fewer children ages 6-8

getting no homework and more children

receiving some homework. Homework has

not become overwhelming. Data from NAEP

indicate that more than 80% of children at

age 9 have less than an hour of homework

per day. Only one out of twenty have more

than two hours. For students in middle

school and high school, the amount of home-

work has remained stable—and hardly bur-

densome. According to UCLA’s annual

survey of college freshmen, the homework

required in the senior year of high school is

particularly light. And it has been declining

since the late 1980s. According to the TIMSS

international survey, American high school

students have one of the lightest homework

loads in the world. Parents are not con-

cerned about too much homework.

According to a study by the Public Agenda

Foundation, only one of ten parents want less

homework for their children, 25% want

more, and two-thirds describe the home

work load as about right.

OK, we can all relax; kids aren’t being

worked to death. But what about parents with

legitimate concerns? What should they make

of the homework controversy? The problem

75.8

Socializing with friends

0%

10%

20%

30%

40%

60%

70%

80%

50%58.3

Working (for pay)

49.9

Exercise or sports

33.4

Studying/ homework

26.0

Watching TV

25.1

Partying

14.0

Student clubs/groups

Studying does not dominate the lives of high school seniors.

Socializing, working, and sports command more time than studying.

Students spending more than five hours per week on activities in their last year of high school

Source: Linda Sax, J.A. Lindholm, A.W. Astin, W.S. Korn, and K.M. Mahoney, The American Freshman: National Norms for Fall 2002, (Higher Education Research Institute, UCLA, 2002).

(percentage)

Fig

2-3

According to UCLA’s

annual survey of

college freshmen, the

homework required

in the senior year of

high school is particu-

larly light. And it has

been declining since

the late 1980s.


of too much homework may be exaggerated

in the press, but it can exist. On any particular

school night, the most diligent student in the

world may come home with an unmanageable

amount of homework. Or the most reasonable

teacher in the world may assign an unreason-

able amount. Teachers are human and they

make mistakes. What can parents learn from

the homework controversy that will help

them monitor their children’s workloads? The

following suggestions make the most sense.

1. Take anti-homework articles

with a grain of salt.

Opposition to homework has been a main-

stay of progressive education for the past

100 years. Edward Bok, editor of The Ladies

Home Journal, launched a campaign against

homework in 1900. Bok argued that chil-

dren under the age of 15 shouldn’t be

assigned any homework.21 After an inten-

sive media campaign in the early twentieth

century, progressives managed to persuade

lawmakers in one state, California, to ban

homework. Progressives charged that staying

indoors, reading books, and engaging in

other intellectual activities were unhealthy

for children.

Progressives revere children’s natural

development, which they believe unfolds

through play and self-guided exploration.

Homework is work. It interferes with play. It

is assigned by an adult. Progressives are hos-

tile to homework because “children need

time to be children.” Progressives also

believe that education is best left to experts.

Children who take home schoolwork might

be taught “incorrectly” by their parents. And

parents who scrutinize their children’s home-

work will have a good idea of what’s being

25%Too little

10%Too much

2%Don’t know

64%About right

Parents are not concerned about too much homework.

Parents of K-12 students were asked what they thought about their children’s homework load.

Fig

2-4

Source: Public Agenda Foundation, Questionnaire and Full Survey Results, National Poll of Parents of Public School Students, (Public Agenda Foundation, September 2000).


© 2003 People, reprinted by permission.



taught at school, an invitation for parental

meddling in school affairs.22

2. Follow the PTA guidelines

on homework.

Based on the work of Harris Cooper at the

University of Missouri, the Parent-Teacher

Association offers as a rule of thumb

10 minutes of homework per grade—

for example, 30 minutes a night for third

graders, 60 minutes for sixth graders,

90 minutes for ninth graders. That seems

reasonable. However…

3. Understand that homework varies.

Homework fluctuates from day to day.

Teachers assign different amounts depend-

ing on what is being studied. Kids have

days when they are extremely productive

and get a lot of work done at school. Other

days they are not productive, and the work

comes home. Homework also naturally

varies from child to child. Some kids use

every spare minute at school to complete

their work. They take home virtually noth-

ing. Other kids enjoy socializing at school.

Companions are sitting next to them when

they are at school, not when they are at

home in the evenings. Others simply prefer

to work at a leisurely pace. These students

take home full backpacks. Plus, once at

home, some kids spend an hour just to get

ready to do homework—not actually to do

it, mind you, but to get ready—sharpening

a pencil, getting the lighting just right,

making sure that the pet dog is in his prop-

er place, adjusting the mirror on the wall,

pouring a glass of water, sharpening a back-

up pencil...and on and on. Some kids get

right to work and finish quickly. The

upshot is that individual circumstances

must be considered when hearing about

huge amounts of homework. Children

do not all possess the same study habits.

4. If a homework problem exists,

solutions should come from parents and

teachers, not policy interventions.

This recommendation flows from the

previous point. Homework loads vary by

student. Even in the same family, children

experience homework differently. If parents

believe their child has too much home-

work, they should talk to the teacher, or

teachers in the case of high school. It is

a conceptual mistake to try to standardize

school assignments in units of time. If

the page you are currently reading were

assigned to thirty high school students

to read, they would take varying amounts

of time to complete the task. Teachers

should try to standardize the content of

schoolwork, not the time to complete it.

Only teachers and parents working together

can assess the individual circumstances

regarding homework and determine

whether children’s homework loads are

appropriate. Making such decisions away

from classrooms is inappropriate. For

school districts to place limits on home-

work, such as the Piscataway District in

New Jersey recently did, seems unwise.23

For state legislatures to pass laws on home-

work seems downright silly.

Only teachers and

parents working together

can assess the individ-

ual circumstances

regarding homework

and determine whether

children’s homework

loads are appropriate.


Part

IIICHARTER SCHOOLS:ACHIEVEMENT,ACCOUNTABILITY,AND THE ROLE OF EXPERTISE


Part III Charter Schools: Achievement, Accountability, and the Role of Expertise

The study combined data over three years

to form a composite achievement score

for each school. A three-year composite

smoothes out some of the year to year

fluctuations in test scores. The study dis-

covered that charters score –.24 z-scores or

approximately one-fourth standard

deviation below regular public schools

with similar demographic profiles.

As was cautioned last year, examining

test score levels is not conclusive in evaluat-

ing the effectiveness of charter schooling.

Many parents pull their children out of

regular public schools and send them to

charter schools because of unhappiness

with how much their children are learning.

Researchers call this a “selection effect.”

When low achievers select charter schools,

the effect depresses charter school test

scores irrespective of the quality of educa-

tion charters are providing. Charter schools

may also experience a positive selection

effect by attracting parents who are extraor-

dinarily concerned about their children’s

success at school and are highly motivated

to intervene on their children’s behalf.

Selection effects limit the ability of analysts

to make causal inferences from charter

schools’ test scores. That said, an analysis of

scores does give a reasonable estimate of

how well students in charter schools—or

any other type of school—are functioning

academically at a given point in time. In

discussing charter schools and academic

L AST YEAR’S BROWN CENTER REPORT INCLUDED AN ANALYSIS

of charter school test scores from 1999 through 2001. Following on

the heels of several state-specific studies, the study was the first to assess

charter school achievement nationwide. Brown Center researchers computed

z-scores for charter schools—indexing charter schools’ test scores relative

to the mean and standard deviation of test scores within each state—and

then examined the z-scores nationally. This strategy, although possessing its own

limitations, gets around the problem that states use different achievement

tests and report scores in different metrics.24


achievement, test score levels are a starting

point, not the final word.25

This year’s report offers a follow up

study with several new elements. (Last

year’s sample will be referred to as the 1999

cohort and this year’s as the 2000 cohort.)

The 2000 cohort consists of all charter

schools that were open in 2000 and have

three years of test data—through 2002—in

state databases. The sample size grew from

376 schools in the 1999 cohort to 569

schools in the 2000 cohort. Last year’s anal-

ysis focused on test score levels, with math

and reading scores combined for each year.

This year’s study includes an analysis of

both levels and test score changes from

2000 to 2002. The latter will be referred to

as “gains” or “gain scores” even though loss-

es may occur. Gain scores indicate whether

students in a particular year are scoring

higher or lower than students at the same

school in a previous year.

Both levels and gains play important

roles in accountability systems. The federal

government’s No Child Left Behind Act

requires states to use a school’s achieve-

ment level from a starting year (or “base-

line”) and calculate the annual gains

needed for 100% of students to reach aca-

demic proficiency by 2014. This statistic

becomes the standard by which schools

demonstrate Adequate Yearly Progress

(AYP), the key accountability measure in

the act. Adequate progress is not the same

for all schools. Schools with low baseline

scores have more improvement to make to

get to 100% proficiency and as a conse-

quence have higher AYP targets; those with

higher baseline scores do not have to

improve as much and have lower AYP tar-

gets. Schools failing to meet their AYP tar-

gets face sanctions.

The main lesson from last year’s study

of charter schools was that, compared to

regular public schools, charter schools have

more ground to make up and as a conse-

quence are more likely to be subjected to

sanctions in accountability systems. States

issue annual lists of failing schools facing

possible sanctions. Below, we examine how

often charters appear on the 2003 lists.

This year’s study concludes by scruti-

nizing two forms of expertise found in

charter schools. Conversion charters—

schools that have converted from regular

public schools to charter status—constitute

about one-third of the charter schools

in California. The test scores of conversion

charters in California are analyzed. When

a regular public school converts to a

charter school, the expertise of teachers and

administrators who stay at the school car-

ries over. Conversion schools are funda-

mentally different from charters that start

from scratch.26

A second form of expertise can be

found in educational management

organizations (EMOs). A 2003 study by

the Education Policy Studies Laboratory

at Arizona State University identified 47

companies managing 417 schools in 24

states. Chancellor Beacon Academies,

Edison Schools, The Leona Group, and

National Heritage Academies are among the

most prominent EMOs.27 These firms usu-

ally operate for profit. They either

contract to manage charter schools or hold

charters outright, with authority over

hiring and firing personnel, developing cur-

ricula, and providing instruction.

Proponents of EMOs claim that they inject

new blood into the educational system

by bringing entrepreneurs onto school cam-

puses who may otherwise not be involved

with public education. The study compares

test scores of EMO-managed charter

schools to scores of other charters and regu-

lar public schools.

Schools that have converted from regular public schools to charter schools.

A firm dedicated to operatingschools or districts,usually on a for-profit basis.

EMO(educational management organization)

Conversion charters



Conversions are more

likely to be located in

urban areas than other

charters or regular

public schools and more

likely to serve black

students and students

in poverty.

A caveat. The study is not meant to

serve as an evaluation of charter schools.

The most reliable way to do that is through

a randomized field trial. The fact that seats

in over-subscribed charter schools are

granted by lottery allows for the creation of

randomly selected treatment and control

groups, thereby making it less likely that

selection effects are tainting achievement

outcomes. Randomized field trials are

extremely expensive to conduct, however,

and as of the summer of 2003, one has not

yet been conducted with charter schools.

Charter School AchievementThe study focuses on ten states with signifi-

cant numbers of charter schools. Technical

details can be found in the 2002 Brown

Center Report. Here are the basics. The ten

states administer several different tests to

measure student achievement, so analyzing

raw test scores is only possible within a

single state. To examine the performance of

schools across states, scores must be

expressed in the same metric. Brown Center

researchers converted raw scores to z-scores

for the 569 charter schools and 25,614

regular public schools in the ten states.

The z-score of a school reveals the

school’s performance relative to the state

average—negative scores indicating how far

below average in standard deviation units,

positive scores, how far above average. By

definition, in any particular year, the average

z-score for all schools in a state is 0.00. After

adjusting the z-scores for poverty and racial

composition, we found the average z-score

for the 569 charters in the study. The average

score of –.31 in 2002 means that charter

schools score significantly lower than regular

public schools in their own states with com-

parable demographic characteristics.

In terms of percentile ranks, a z-score

of –.31 is equal to the 38th percentile.

Percentiles range from 1 to 99, with 50 being

the median score, the point at which half

rank above and half rank below. The 50th

percentile is also the average score. A score

at the 38th percentile indicates that about

62% of schools with similar demographic

characteristics have higher test scores than

the average charter school.28

Table 3-1 also reveals that charter

schools made steady improvements in

scores each year, from –.53 in 2000, to –.40

in 2001, to –.31 in 2002. The positive gains

registered by the 2000 cohort of charter

schools are a significant change from last

year’s study. Gains registered from 1999 to

2001 for the 1999 cohort were insignificant

from zero (+.02). Charters’ test scores

rose, but by the same amount as regular

public schools. Test scores for the 2000

cohort of charters, on the other hand, are

rising sharply and the +.22 z-score gain

exceeds the gains made by all public schools

(which also is 0.00 by definition).

What is different about the two

cohorts? The new charters that opened in

2000 did not experience the negative “two

year effect” described in last year’s study. In

the 1999 cohort, test scores were depressed

for the first two years that new charters

were open. In the current study, the 126

charter schools that opened for the first time

in 2000 gained .18 z-scores by 2002.

Charters and AccountabilityHow are charter schools faring under state

accountability systems? Table 3-2 displays

the proportion of charters on state lists of

failing schools, based on lists that were

available on state websites in August, 2003.

Charters are over-represented. The table

displays data for the ten states in the study.

Of 569 charters, 140 schools (24.6%) were

failing. Among all public schools in the

same ten states, 21.3% were failing. This is


the second year in a row that we have found

charters over-represented on state failure

lists. The comparable figures in 2002

were 18.6% for charters and 12.3% for all

public schools, so the over-representation

is shrinking.29 A cautionary word. States

may change policies regarding how failing

schools are identified, and a single state can

strongly influence the national average.

Pennsylvania ratcheted up its standards in

2003. In the cohort of charter schools

studied last year, none from Pennsylvania

were failing. This year, 25 out of 32 charters

appear on the state’s failing schools list.

The statewide failure rate for all public

schools in Pennsylvania increased from

8.4% to 40.3%.

As noted above, failure in account-

ability systems is predicated on two aspects

of test scores—level and gain. For the 1999

cohort, findings were negative in terms

of levels and neutral in terms of gains. So

it is not surprising that a larger share of

charters than regular public schools were

found on lists of failing schools. The 2000

cohort fares a little better. The news is

again negative in terms of levels, but the

significant gains that charters have made

on achievement tests have kept many of

them off failing schools lists.

California ConversionsCalifornia has the largest proportion of

conversion charter schools of any state in

the nation. After the state’s 1992 charter

law was enacted, most of the schools char-

tered in the first few years were conver-

sions. Since then start-ups have dominated

as new, independent schools were launched

by parents, teachers, concerned citizens,

and entrepreneurs. Conversions constituted

about one-third of the state’s charters oper-

ating in 2000.30 Schools may petition local

school boards for permission to convert to

charter status. If the petition is denied,

schools may appeal to the local county

board of education. Unlike several states,

California prohibits private schools from

becoming charters. Conversions are freed

from district regulations beyond those

that are negotiated with districts during the

conversion process and stipulated in the

charters. Some charter schools contract with

former districts to provide food services,

transportation, or administrative services.31

Charter School Achievement(scores expressed as adjusted z-scores, N=569)

Z-score

2000

–0.53*(0.05)

2001

–0.40*(0.05)

2002

–0.31*(0.05)

2000-2002Z-score Gain

+0.22*(0.04)

Table

3-1


NOTE: Z-scores adjusted for poverty and racial composition. Standard error in parentheses.

Number of Failing Charter Schools(August 2003)

State

Arizona

California

Colorado

Florida

Massachusetts

Michigan

Minnesota

Pennsylvania

Texas

Wisconsin

Total

Schools in Study

71

132

51

59

26

111

21

32

48

18

569

NumberFailing

8

69

1

0

9

13

7

25

6

2

140

PercentFailing

11.2%

52.2%

2.0%

0.0%

34.6%

11.7%

33.3%

78.1%

12.5%

11.1%

24.6%

Schools in State

1,489

8,238

1,516

2,616

1,858

3,512

1,969

3,172

6,894

2,065

33,329

NumberFailing

346

3,715

86

10

209

121

265

1,279

1,000

72

7,103

Table

3-2

NOTE: Each state has its own criteria, based on individual state tests, for determining whether or not aschool is failing.

Source: Data compiled from respective state department of education websites.

PercentFailing

23.2%

45.1%

5.7%

0.3%

11.2%

3.4%

13.4%

40.3%

14.5%

3.5%

21.3%

Charter Schools All Public Schools



Table 3-3 displays characteristics of

the conversions. They are much larger than

other charters. With a median enrollment

of 633 students, conversions resemble

California’s traditional public schools in

size. Conversions are more likely to be

located in urban areas than other charters

or regular public schools and more likely

to serve black students and students in

poverty. Although conversions are only

about one-third of charters in California,

they constitute exactly one-half of the

current study’s 2002 cohort of charters.

Conversions are more likely to show up

in this sample because they tend to be

older, larger charters and therefore more

likely to have met the requirements for

the study (i.e., have three consecutive

years of test data).

Table 3-4 exhibits the z-scores of

the different types of schools, adjusted for

racial composition and poverty. Conversions’

test scores are higher than those of other

charters and regular public schools. In 2002,

conversions scored .20 z-scores above the

average school in California with similar

demographic characteristics. This is equal to

about the 58th percentile. Start-up charters

scored significantly below average (–.40),

at approximately the 34th percentile.

Charters as a whole scored –.10, which is

statistically indistinguishable from 0.00, the

average for California’s regular public

schools. The change in scores from 2000 to

2002 is essentially the same for conversions,

start-ups, and regular public schools.

This raises an important drawback of

using z-scores to track changes in test

scores. Z-scores are relative scores. They

tell us how one competitor is doing in a

race—outdistancing the others or falling far

behind—but they do not reveal whether the

participants as a group are running or walk-

ing. In other words, by themselves z-scores

cannot tell us whether test scores are going

up, down, or sideways. In Table 3-4, they

only tell us that—whatever the direction—

the scores of conversion charters, start-up

charters, and regular public schools have

changed approximately the same amount.

None of the three types of schools stand out

as having improved or declined more than

the others.

David Rogosa of Stanford University

has written extensively on achievement

in California schools. Rogosa cautions that

making comparisons based on residuals

from regression equations, as the z-scores

presented here do to control for demo-

graphic characteristics, may produce mis-

leading results—especially when applied

to changes in test scores. He points out that

researchers often imply that they are

answering the question, “How much would

the school have changed if all schools had

started out equal? rather than the answer-

able question, how much did the school

improve?”32 As useful as regression analy-

sis may be, it cannot answer questions

about imagined conditions. And packing

several variables into regression equations

in order to isolate the effects that are of pri-

mary interest may increase the chances of

getting the story wrong.

As a check on the preceding analysis,

we examined the national percentile rank-

ings (NPRs) of conversion charters on the

Stanford Achievement Test (SAT-9) from

2000-2002 and the gains schools registered

Granting charter status to existing schools, in other words,

has probably functioned more as a reward for demonstrating

success than as a turn-around or intervention strategy for

low performance.


in NPRs during the same period. Table 3-5

presents the data. In 2002, charters scored

at about the 50th percentile. Regular

public schools scored slightly higher, at the

52.7 percentile. From 2000 to 2002, the

different types of schools all gained 3 to 4

percentile points.

These results clarify the findings pre-

sented above. Let’s sum them up. From

2000 to 2002, test scores of charter schools

in California were lower than those of regu-

lar public schools. During the same period,

scores rose by about the same amount in

California’s charter and regular public

schools—3 to 4 percentile points on the

SAT-9. Among the charters, conversions

and start-ups scored at about the same

level, and their scores increased about the

same amount from 2000 to 2002. Any anal-

ysis that controls for demographic charac-

teristics will find conversions scoring

significantly higher than both start-ups and

regular public schools. This is because con-

versions serve large numbers of students

with socioeconomic characteristics that are

correlated with low test scores. Gain scores

look about the same for conversions and

start-ups when demographic controls are

employed with the data.33

Many conversion charters are produc-

ing average test scores with populations of

children historically associated with low

test scores. These schools may be doing

something that is worth identifying and dis-

seminating to others. The phenomenon

may also be related to how conventional

public schools convert to charter status.

Local school boards grant charters in

California. They are probably reluctant to

approve charter petitions coming from a

district school unless the petitioners have a

solid track record. Granting charter status

to existing schools, in other words, has

probably functioned more as a reward for

Characteristics of California Conversion Charter Schools(2002)

Enrollment (median)

Poverty

White

Black

Hispanic

Asian

Urban

Suburban

Rural

Conversion Charters(N=66)

633

58%

33%

19%

41%

6%

55%

36%

8%

Start-up Charters(N=66)

234

42%

57%

13%

23%

4%

40%

38%

22%

California Regular Public Schools

(N=6,623)

669

49%

38%

8%

41%

11%

34%

55%

12%

Table

3-3

NOTE: Mean enrollments: Conversion = 746, Start-up = 487, CA Regular Public School = 837

California Charter School Achievement(scores expressed as adjusted z-scores)

Regular Public Schools(N=6,623)

All Charters(N=132)



2000

0.00(.01)

–0.08(.07)

0.24*(.08)

–0.39*(.11)

2001

0.00(.01)

–0.07(.07)

0.24*(.07)

–0.39*(.12)

2002

0.00(.01)

–0.10(.08)

0.20*(.08)

–0.40*(.12)

2000-2002 Z-score Gain

0.00(.01)

–0.02(.06)

–0.04(.07)

–0.01(.08)

Table

3-4


NOTE: Conversion charter scores are statistically significantly different (p < .05) from start-up charterscores in all years. Z-scores adjusted for poverty and racial competition. Standard error in parentheses.

California Charter School Achievement(scores expressed as unadjusted national percentile ranks)


All Charters(N=132)



2000

49.2(0.2)

46.5(1.7)

46.3(2.5)

46.7(2.3)

2001

51.2(0.2)

48.9(1.7)

48.6(2.5)

49.1(2.3)

2002

52.7(0.2)

50.1(1.7)

50.2(2.4)

50.0(2.3)

2000-2002NPR Gain

+3.5(0.1)

+3.6(0.9)

+3.9(1.3)

+3.3(1.3)

Table

3-5

NOTE: SAT-9 scores used to measure achievement. Standard error in parentheses.



demonstrating success than as a turn-

around or intervention strategy for low per-

formance. This is important for researchers

to note. The selection factors that plague

the evaluation of charter school achieve-

ment may play out differently with conver-

sions, especially if a school has

institutional prestige that precedes its

charter status.

Educational ManagementOrganizations (EMOs)EMOs operate 90 charter schools in the

study’s sample of ten states (see Table 3-6).

Michigan dominates the EMO landscape,

with 62 of the 90 schools located there.

Like California’s conversions, the EMO

schools are larger than the average charter.

Compared to both other charters and regu-

lar public schools, they serve a larger pro-

portion of black children and children in

poverty and are more likely to be located in

urban communities. They are less likely to

serve Hispanic students, which is due to

relatively few EMOs operating in states

such as California, Florida, and Texas with

large Hispanic populations.

Charters run by EMOs typically

target low-achieving students. Table 3-7

reveals that EMOs scored significantly

below non-EMO charters and further

yet below regular public schools, with

controls in place to compare schools with

similar demographic characteristics. The

EMO charters’ z-score in 2000 (–1.00)

was equal to about the 16th percentile.

In contrast to conversion charters, EMO-

operated charters apparently are not

brought in to manage schools in recogni-

tion of schools’ success, but to intervene in

the case of school failure.

EMOs are controversial. Many people

are hostile to the notion of firms profiting

financially from public schools. When

charter authorizers consider approving

an EMO-operated charter, it often gener-

ates intense political opposition. The test

score data suggest that in situations where

student achievement is extremely low,

the political opposition to EMOs may

be weakened.34

EMO-operated charters registered

solid gains in test scores from 2000 to 2002

(see last column, Table 3-7), significantly

out-gaining non-EMO charter schools with

similar demographic profiles. In 2002, the

test scores of the 90 EMO charters in the

study had moved up to approximately the

28th percentile (z-score of –.58). To run a

check on the analysis using raw test scores,

we analyzed the scores of EMO-operated

charters in Michigan on the Michigan

Educational Assessment Program (MEAP).

The z-score and raw score treatments yield-

ed essentially the same results. EMO-operat-

ed charters score significantly lower than

regular public schools and non-EMO char-

ters but registered greater test score gains

from 2000 to 2002.

Summary and ConclusionThis section of the report analyzed test

score data of charter schools. As was found

last year in the first segment of the study,

test scores in charter schools lag behind the

scores of regular public schools. Charter

schools in this year’s study—contrary to the

charters in last year’s—registered significant

gains in test scores from 2000 to 2002. On

the lists of failing schools that states com-

piled in both 2002 and 2003, charter

schools are over-represented.

Two forms of expertise were exam-

ined by analyzing achievement data from

conversion charters in California and EMO-

operated charters in several states. In 2002,

the conversion charters scored at approxi-

mately the 50th percentile on reading and

Future research should

focus on identifying the

broadest range of

policies and practices

that are capable of

creating achievement

successes out of

achievement failures.


math tests. From 2000 to 2002, they made

gains comparable to those of start-up char-

ters and regular public schools. Conversions

produced average test scores despite a

demographic profile that is usually correlat-

ed with low scores.

A recent study by Bruce Fuller at the

University of California, Berkeley found

that charter schools have fewer fully certi-

fied teachers than regular public schools.35

Conversion charters are an exception, with

their teachers comparable to those of regu-

lar public schools in experience and certifi-

cation. Can the conversions’ solid test

scores be attributed to teachers’ expertise?

Perhaps, but not necessarily. The test scores

of non-conversion charters are improving

just as much. And, as noted above, neither

Fuller’s nor this study can control for selec-

tion effects—in terms of which schools

were allowed to convert or the students

who chose to attend conversions and start-

ups. Much more research is needed com-

paring the performance of conversion and

start-up charters.

The test scores of charters operated

by educational management organizations

were also examined. Compared to regular

public schools and to other charters serving

students with similar socioeconomic

characteristics, EMO-operated charters

have much lower test scores. This is almost

certainly because companies have been

chartered to run already low-performing

schools or to create new schools serving

low-performing students. Gains made from

2000 to 2002 have been significantly larger

than those of both non-EMO charters

and regular public schools. Again, causality

cannot be determined conclusively. However,

importing management expertise from the

private sector should be explored further

as an option for improving low-achieving

schools. The results presented here are

quite positive for EMOs. The findings raise

doubts about a strain of thought in the

charter school movement: that anyone can

successfully start and operate a school if

he or she merely possesses abundant energy

and a love of children. In the extreme, this

form of romanticized amateurism dismisses

the importance of educational expertise.

That appears unwise. Future research should

focus on identifying the broadest range

of policies and practices that are capable

of creating achievement successes out of

achievement failures.

EMO Charter Achievement(scores expressed as adjusted z-scores)


All Charters(N=569)

EMO Charter(N=90)

Non-EMO Charter (N=479)

2000

0.01(.01)

–0.53*(.05)

–1.00*(.10)

–0.44*(.06)

2001

0.01(.01)

–0.40*(.05)

–0.69*(.09)

–0.35*(.06)

2002

0.01(.01)

–0.31*(.05)

–0.58*(.08)

–0.26*(.05)

2000-2002Z-score Gain

0.00(.01)

+0.22*(.04)

+0.41*(.08)

+0.18*(.05)

Table

3-7


NOTE: Z-scores adjusted for poverty and racial composition. Standard error in parentheses.

Characteristics of EMO Charter Schools(2002)

Enrollment (median)

Poverty

White

Black

Hispanic

Asian

Urban

Suburban

Rural

EMO(N=90)

498

53%

43%

41%

14%

1%

45%

41%

15%

Non-EMO(N=479)

248

47%

51%

25%

19%

3%

52%

33%

15%

Regular Public School(N=25,614)

546

42%

57%

12%

25%

5%

30%

40%

29%

Table

3-6

NOTE: Mean enrollments are: EMO = 507, Non-EMO = 369, Regular Public School = 664


1 Approximating one year’s growth on NAEP is explained in the 2000 Brown Center Report on AmericanEducation, p. 7.

2 Richard L. Allington, “Does State and FederalReading Policymaking Matter?” in The GreatCurriculum Debate: How Should We Teach Reading andMath? edited by Tom Loveless, (Brookings InstitutionPress, 2001), pp. 268-298.

3 Linda Dager Wilson and Rolf K. Blank, ImprovingMathematics Education Using Results from NAEP andTIMSS, (Council of Chief State School Officers, 1999).

4 National Commission on the High School Senior Year, The Lost Opportunity of Senior Year: Finding a Better Way, (National Commission on the High SchoolSenior Year, 2001).

5 The counts were stable across years.

6 School funding data reported in “Public Elementaryand Secondary Students, Schools, Pupil/Teacher Ratios, and Finances, by Type of Locale: 1998 and1999” on the National Center for Education Statistics“Navigating Resources for Rural Schools” website(http://nces.ed.gov/surveys/ruraled). Cost of livingfigures from Mark Nord, “Does It Cost Less to Live in Rural Areas? Evidence from New Data on FoodSecurity and Hunger,” Rural Sociology, vol.65, no.1 (March, 2000), pp. 104-125.

7 The Rural School and Community Trust, Why RuralMatters 2003, (The Rural School and CommunityTrust, 2003).

8 “Navigating Resources for Rural Schools,” National Center for Education Statistics website:http://nces.ed.gov/surveys/ruraled

9 Dana Markow and Marc Scheer, The MetLife Surveyof The American Teacher 2002: Student Life: School,Home, and Community,(MetLife, Inc., 2002); LowellRose and Alec Gallup, “The 34th Annual Phi DeltaKappa/Gallup Poll Of the Public’s Attitudes Toward thePublic Schools,” Phi Delta Kappan supplement(September, 2002).

10 Richard Ingersoll, “Teacher Turnover and TeacherShortages: An Organizational Analysis,” AmericanEducational Research Journal, vol. 38 no. 3 (Fall, 2001),pp. 499-534; also see U.S. Department of Education,National Center for Education Statistics, Digest of Education Statistics 2001, (NCES 2002-130, U.S.Government Printing Office, 2002), Table 75.

11 Mihaly Csikszentmihalyi and Barbara Schneider,Becoming Adult: How Teenagers Prepare for the World ofWork, (Basic Books, 2001). Also see U.S. Department of Education, National Center for Education Statistics,Digest of Education Statistics 2001, (NCES 2002-130,U.S. Government Printing Office, 2002), Table 186.

12 CNN.com, “Alaskan Schools Mired in RusticConditions,” May 14, 2003.

13 Harris Cooper, The Battle Over Homework:Common Ground for Administrators, Teachers andParents, (Corwin Press, 2001).

14 Sandra L. Hofferth and John F. Sandberg,“Changes in American Children's Use of Time, 1981-1997,” Children at the Millennium: Where Have We ComeFrom, Where Are We Going? edited by S. Hofferth and T.Owens, (Elsevier Science Ltd., 2001), pp. 193-229.

15 The website for CBS Evening News lists a story forDecember 10, 2002, “Study Says Kids’ HomeworkLoad Bigger,” which includes the following, “Since1981, homework has stayed about the same for highschool students—but for 6- to 9-year olds, it’s tripled.”The report does not list a source. The Michigan studydid not include high school students, but the CBSreport’s reference to 1981 and ages 6 to 9, reasonablyclose to the ages 6 to 8 category in the study, indicatesit probably was the source. The website forCNNfyi.com includes an AP story, July 25, 2001,“Homework Hours Tripled Since 1980,” stating thefollowing, “One University of Michigan study suggeststhat young children are seeing up to three times asmuch homework as children did 20 years ago.”

16 Sandra L. Hofferth and John F. Sandberg,“Changes in American Children's Use of Time, 1981-1997,” Children at the Millennium: Where Have We ComeFrom, Where Are We Going? edited by S. Hofferth and T.Owens, (Elsevier Science Ltd., 2001), pp. 218, 226.

17 NAEP Reading 1999 Long-term trend summarydata. Other NAEP surveys ask for a daily estimateusing different response categories. The results lead toessentially the same conclusions reached here.

18 Maryellen Harmon, Teresa A. Smith, Michael O.Martin, Dana L. Kelly, Albert E. Beaton, Ina V.S. Mullis,Eugenio J. Gonzalez, and Graham Orpwood,Performance Assessment in IEA’s Third InternationalMathematics and Science Study, (The InternationalStudy Center at Boston College, 1997).

19 Linda Sax, Jennifer Lindholm, Alexander Astin,William Kirn, and Kathryn Mahoney, The AmericanFreshman: National Norms for Fall 2002, (HigherEducation Research Institute, UCLA, 2002).

20 Public Agenda Foundation, Questionnaire and FullSurvey Results, National Poll of Parents of Public SchoolStudents, (Public Agenda Foundation, 2000).

21 Diane Ravitch, Left Back: A Century of Failed SchoolReforms, (Simon & Schuster, 2000), p. 90.

22 Brian Gill and S.L. Schlossman, “ ‘A Sin AgainstChildhood’: Progressive Education and the Crusade to Abolish Homework, 1897- 1941,” American Journalof Education, vol. 105, no. 1 (November, 1996), pp. 27-66; Gill and Schlossman, “Parents and thePolitics of Homework: Some Historical Perspectives,”Teachers College Record, vol. 105, no. 5 (June, 2003),pp. 846-871.

23 CNN.com “Restrictions on Homework Put NewJersey School District in Spotlight,” October 23, 2000.

24 Some limitations: z-scores do not tell whether testscores are rising or falling statewide; and they mayimply that one standard deviation in one state is equalto one SD in another.

25 Gains provide a better control if the basis onwhich selection occurs remains stable from the startand end point of the time period for which the gain iscalculated.

26 For the challenges facing start from scratch char-ters, see Tom Loveless and Claudia Jasin, “Startingfrom Scratch: Organizational and Political ObstaclesFacing Charter Schools,” Education AdministrationQuarterly, vol. 34, no. 1 (February, 1998), pp. 9-30.

27 Alex Molnar, Glen Wilson, and Daniel Allen,Profiles of For-Profit Education Management Companies:Fifth Annual Report 2002-2003, (Education PolicyStudies Laboratory, Arizona State University, 2003)

28 Last year’s calculations were weighted by enroll-ment to give larger schools more influence than small-er schools. For the sake of simplicity, we did notweight the means reported this year. Two reasons: (1)The three-year composite already smoothes out fluctuations that are endemic to small schools’ testscores. (2) When we ran the analyses using weights,the results were not significantly different from theoutcomes reported here.

29 Failing lists for 2002 are analyzed in Tom Loveless“Charter School Accountability and Achievement,” inLeave No Child Behind: The Politics and Practice of SchoolAccountability, edited by Paul Peterson and MartinWest, (Brookings Institution Press, forthcoming).

30 Priscilla Wohlstetter, Noelle C. Griffin, andDerrick Chau “Charter Schools in California,” in TheCharter School Landscape, edited by Sandra Vergari,(University of Pittsburgh Press, 2002), pp. 32-53.

31 Web information on California conversions isavailable at: http://www.cde.ca.gov/charter/about.html

32 David Rogosa, “Student Progress in California Charter Schools, 1999-2002,” (California Department of Education, 2003).

33 Rogosa finds, and our data confirm, that gainscores have a modestly positive association with indica-tors of poverty. This is enormously intriguing, and wehope to explore it in greater depth in the future. SeeRogosa, 2003, for his discussion.

34 For an example of strong public backlash againstEMO-operated schools, note the case of EdisonSchools’ withdrawal from San Francisco in JulianGuthrie, “S.F. Schools Vote to End Edison Compact,”San Francisco Chronicle, June 29, 2001.

35 See the working paper by Bruce Fuller, MarytzaGawlik, Emlel Kuboyama Gonzales, Sandra Park, andGordon Gibbings, “Charter Schools and Inequality:National Disparities in Funding, Teacher Quality, andStudent Support,” (Policy Analysis for CaliforniaEducation, 2003).

ENDNOTES

Des

igne

d by

Lip

man

Hea

rne,

Chi

cago

/Was

hing

ton

THE BROOKINGSINSTITUTION

STROBE TALBOTTPresident

CAROL GRAHAMVice President and Director of Governance Studies

BROWN CENTER STAFF

TOM LOVELESSSenior Fellow and Director

PAUL T. HILLNon-resident Senior Fellow

DIANE RAVITCHNon-resident Senior Fellow

PAUL DIPERNASenior Research Analyst and Brown Center Manager

JOHN COUGHLANResearch Assistant

KATHLEEN ELLIOTT YINUGCoordinator Brookings Papers on Education Policy

ADVISORY & REVIEW BOARD

CHRISTOPHER N. AVERYHarvard University

PAUL T. HILLUniversity of Washington

THOMAS J. KANEUniversity of California-Los Angeles

DIANE RAVITCHNew York University

BARBARA SCHNEIDERUniversity of Chicago

RESEARCH VERIFIER

JIMMY KIMHarvard University

Views expressed in this report are solely those of the author.

1775 Massachusetts Avenue, NW • Washington, DC 20036Tel: 202-797-6000 • Fax: 202-797-6004

www.brookings.edu

Tel: 202-797-6426 • Fax: 202-797-2968www.brookings.edu/browncenter

Documents

Brown Center Report 2003 - Brookings2016/06/10 · This is the fourth edition of the Brown Center Report on American Education. The report premiered in the fall of 2000, as the presidential